LID score v2

[hhhaaahhhaa]

Fully decouple LID scoring and TER scoring, and parse LID directly from the text file. The previous version is bugged when dealing with [org_jpn].

[codecov]

Codecov Report

Merging #4983 (cafe59b) into master (b5b2b11) will increase coverage by 0.10%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #4983      +/-   ##
==========================================
+ Coverage   76.91%   77.01%   +0.10%     
==========================================
  Files         606      606              
  Lines       53770    53770              
==========================================
+ Hits        41355    41410      +55     
+ Misses      12415    12360      -55     

Flag	Coverage Δ
test_integration_espnet1	66.33% <ø> (ø)
test_integration_espnet2	47.77% <ø> (+0.77%)	⬆️
test_python	66.85% <ø> (ø)
test_utils	23.35% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
espnet2/tasks/abs_task.py	78.25% <0.00%> (+0.47%)	⬆️
espnet2/tts/espnet_model.py	60.36% <0.00%> (+0.90%)	⬆️
espnet2/hubert/espnet_model.py	52.57% <0.00%> (+1.14%)	⬆️
espnet2/samplers/num_elements_batch_sampler.py	87.50% <0.00%> (+1.25%)	⬆️
espnet2/torch_utils/initialize.py	86.30% <0.00%> (+1.36%)	⬆️
espnet2/train/trainer.py	77.80% <0.00%> (+1.40%)	⬆️
espnet2/schedulers/warmup_lr.py	100.00% <0.00%> (+6.66%)	⬆️
espnet2/tasks/hubert.py	87.31% <0.00%> (+29.85%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more